Route to Reason: Adaptive Routing for LLM and Reasoning Strategy Selection

2025-05-26

AI Summary

Large language models (LLMs) often suffer from excessive computation and “overthinking” in complex reasoning tasks, leading to high token costs and inefficiency. Method: This paper proposes RTR, a dynamic routing framework that adaptively selects the optimal LLM and reasoning strategy based on task difficulty under a fixed computational budget. RTR jointly compresses expert model representations and reasoning strategy embeddings, enabling plug-and-play integration of both black-box and white-box LLMs across four distinct reasoning paradigms. It employs lightweight representation learning and budget-aware decision making to support end-to-end differentiable or non-differentiable adaptation, incorporating seven open-source LLMs. Contribution/Results: Experiments on multiple reasoning benchmarks demonstrate that RTR outperforms the best single-model baseline in accuracy while reducing token consumption by over 60%, achieving a Pareto-optimal trade-off between accuracy and inference efficiency.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Planning, Routing, and Scheduling: Planning with Language Models
• Knowledge Representation and Reasoning: Computational Complexity of Reasoning

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

The inherent capabilities of a language model (LM) and the reasoning strategies it employs jointly determine its performance in reasoning tasks. While test-time scaling is regarded as an effective approach to tackling complex reasoning tasks, it incurs substantial computational costs and often leads to"overthinking", where models become trapped in"thought pitfalls". To address this challenge, we propose Route-To-Reason (RTR), a novel unified routing framework that dynamically allocates both LMs and reasoning strategies according to task difficulty under budget constraints. RTR learns compressed representations of both expert models and reasoning strategies, enabling their joint and adaptive selection at inference time. This method is low-cost, highly flexible, and can be seamlessly extended to arbitrary black-box or white-box models and strategies, achieving true plug-and-play functionality. Extensive experiments across seven open source models and four reasoning strategies demonstrate that RTR achieves an optimal trade-off between accuracy and computational efficiency among all baselines, achieving higher accuracy than the best single model while reducing token usage by over 60%.

Problem

• Dynamic allocation of LMs and reasoning strategies for tasks
• Reducing computational costs and avoiding model overthinking
• Achieving accuracy-efficiency trade-off in reasoning tasks

Innovation

• Dynamic allocation of LMs and reasoning strategies
• Compressed representations for adaptive selection
• Plug-and-play with black-box or white-box models